Enhanced Activation of Mitogen-activated Protein Kinase and Myosin Light Chain Kinase by the Pro33 Polymorphism of Integrin β3 *

Integrin β3 is polymorphic at residue 33 (Leu33 or Pro33), and the Pro33 variant exhibits increased outside-in signaling to focal adhesion kinase and greater actin reorganization. Because focal adhesion kinase activation and an intact cytoskeleton are critical links for integrin-mediated signaling to MAPK, we explored the role of integrin αIIbβ3 in this signaling using Chinese hamster ovary and human kidney 293 cell lines expressing either the Leu33 or Pro33 isoform of β3. Compared with Leu33 cells, Pro33 cells demonstrated substantially greater activation of ERK2 (but not MAPK family members JNK and p38) upon adhesion to immobilized fibrinogen (but not fibronectin) and upon integrin cross-linking. ERK2 activation was mediated through MAPK kinase and required phosphoinositide 3-kinase signaling and an intact actin cytoskeleton. Human platelets and Chinese hamster ovary cells expressing the Pro33 isoform showed enhanced activation of the ERK2 substrate myosin light chain kinase (MLCK) upon adhering to fibrinogen. Furthermore, compared with platelets and cells expressing the Leu33 isoform, the Pro33 variant showed greater α-granule release, clot retraction, and adhesion to fibrinogen under shear stress, and these functional differences were abolished by MLCK and MAPK kinase inhibition. Post-integrin occupancy signaling through MAPK and MLCK after αIIbβ3 cross-linking may explain in part the increased adhesive properties of the Pro33 variant of integrin β3.

Integrin ␤ 3 is polymorphic at residue 33 (Leu 33 or Pro 33 ), and the Pro 33 variant exhibits increased outside-in signaling to focal adhesion kinase and greater actin reorganization. Because focal adhesion kinase activation and an intact cytoskeleton are critical links for integrin-mediated signaling to MAPK, we explored the role of integrin ␣ IIb ␤ 3 in this signaling using Chinese hamster ovary and human kidney 293 cell lines expressing either the Leu 33 or Pro 33 isoform of ␤ 3 . Compared with Leu 33 33 isoform showed enhanced activation of the ERK2 substrate myosin light chain kinase (MLCK) upon adhering to fibrinogen. Furthermore, compared with platelets and cells expressing the Leu 33 isoform, the Pro 33 variant showed greater ␣-granule release, clot retraction, and adhesion to fibrinogen under shear stress, and these functional differences were abolished by MLCK and MAPK kinase inhibition. Post-integrin occupancy signaling through MAPK and MLCK after ␣ IIb ␤ 3 cross-linking may explain in part the increased adhesive properties of the Pro 33 variant of integrin ␤ 3 .

cells, Pro 33 cells demonstrated substantially greater activation of ERK2 (but not MAPK family members JNK and p38) upon adhesion to immobilized fibrinogen (but not fibronectin) and upon integrin cross-linking. ERK2 activation was mediated through MAPK kinase and required phosphoinositide 3-kinase signaling and an intact actin cytoskeleton. Human platelets and Chinese hamster ovary cells expressing the Pro
During the formation of a platelet thrombus, the binding of fibrinogen and von Willebrand factor to integrin ␣ IIb ␤ 3 triggers outside-in signals. These signals promote secretion of ␣and dense granules, the secondary wave of aggregation, filamentous actin formation, cytoskeletal rearrangement, and the formation of platelet membrane vesicles with procoagulant activities (1). Defects in outside-in signaling through ␣ IIb ␤ 3 cause abnormal platelet-mediated clot retraction and aggregation and excessive bleeding in the hereditary disorder Glanzmann's thrombasthenia (2,3) and in mice engineered to express integrin ␤ 3 in which the cytoplasmic tyrosines have been replaced with phenylalanine (4).
A number of molecules and pathways have been identified to be involved in the earliest integrin-mediated outside-in signaling events (5). Several downstream effects and mediators have also been identified, including mitogen-activated protein kinases (MAPKs) and myosin light chain kinases (MLCKs).
MAPKs are a family of serine/threonine kinases activated by diverse extracellular stimuli like growth factors, cytokines, hormones, and other stress factors (6). The MAPK cascade consists of a three-kinase module; MAPK is the most distal, being activated by MEK, which, in turn, is activated by a MEK kinase. Activation of MAPK requires dual phosphorylation at threonine and tyrosine residue in a Thr-X-Tyr motif. There are four known members of the MAPK family: 1) extracellular signal-regulated kinase (p44 ERK1 and p42 ERK2 ), 2) c-Jun Nterminal kinase or stress-activated protein kinase (p46 JNK1 and p55 JNK2 ), 3) the p38 kinase, and 4) ERK5/big MAPK (BMK1). ERK1 and ERK2 are involved in cell growth, proliferation, and adhesion, whereas ERK5 is important for angiogenesis (6,7). JNK and p38 have a role in apoptosis (6). All the MAPK family members except BMK1 have been identified in human platelets, but their function is inadequately understood. MLCK is a Ca 2ϩ /calmodulin-dependent enzyme that phosphorylates Thr 18 and Ser 19 on the regulatory light chain of myosin (8). MLCK contains multiple MAPK consensus phosphorylation sites (PX(S/T)P) (9) and is directly phosphorylated by ERK2 (10). Phosphorylation of myosin light chains (MLCs) by MLCK is a critical regulatory step in myosin function and regulates cell migration, cytoskeletal clustering of integrins, and shape change and secretion in platelets (11).
Integrin ␤ 3 is polymorphic at residue 33 (Leu 33 or Pro 33 ; also known as Pl A1 or Pl A2 , respectively), and the Pro 33 form has been associated with an enhanced adhesive phenotype in cell lines and platelets (12,13) and with acute coronary syndromes in some studies (14). This polymorphism is not rare, and 25% of individuals of Northern European descent express Pro 33 isoforms on their platelets. Compared with Leu 33 , Pro 33 cells exhibit enhanced integrin ␣ IIb ␤ 3 -mediated outside-in signaling to focal adhesion kinase and cytoskeletal-dependent cellular functions like fibrin clot retraction and cell adhesion (12). This raises the possibility that compared with Leu 33 cells, Pro 33 cells can provide more efficient ␣ IIb ␤ 3 outside-in signaling. Evidence supporting a role for focal adhesion kinase and the cytoskeleton in integrin-mediated MAPK activation (15,16) led us to explore whether the Leu-to-Pro substitution at amino acid 33 could modulate ␣ IIb ␤ 3 signaling through MAPK. Using Chinese ham-* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  ster ovary cells (CHO) and 293 cell lines overexpressing equivalent levels of the two isoforms of ␣ IIb ␤ 3 and human platelets, we found that compared with the Leu 33 isoform, the Pro 33 variant of ␤ 3 induced greater outside-in activation of ERK2 and/or MLCK. Inhibition of MLCK and ERK2 activation abolished the increased adhesion to fibrinogen and clot retraction associated with Pro 33 cells, and MLCK inhibition abolished the increased P-selectin secretion in Pro 33 platelets.
Cell Lines and Flow Cytometric Analysis-Stable cell lines overexpressing ␣ IIb ␤ 3 were generated by flow cytometric sorting using monoclonal antibody specific for ␣ IIb ␤ 3 as previously described (12). These included the "vector-only" control CHO cells (designated LK) and CHO cells overexpressing the Leu 33 and Pro 33 isoforms of ␣ IIb ␤ 3 (designated Leu 33 and Pro 33 , respectively). To address concerns about clonal variation that may have occurred in the CHO cell lines, a second set of cell lines was also generated in the human embryonic kidney 293 cell line: line PC/Z, vector-only control; line 293Leu 33 , stably expressing the Leu 33 isoform of ␣ IIb ␤ 3 ; and line 293Pro 33 , stably expressing the Pro 33 isoform of ␣ IIb ␤ 3 . Cell-surface expression of ␣ IIb ␤ 3 on and 293 cells was analyzed by flow cytometry using antibody P2, followed by fluorescein isothiocyanate-labeled anti-mouse antibody (12). The mean channel number corresponding to cell fluorescence intensity was used as a measure of ␣ IIb ␤ 3 surface expression. Assessment of ␣ IIb ␤ 3 expression levels was performed within 24 h of each experiment to assure equivalent expression between the Leu 33 and Pro 33 cell lines.
Adhesion to Immobilized Ligands and Cross-linking of ␣ IIb ␤ 3 Receptors-Cells were grown to 70 -80% confluence and detached using 0.05% trypsin. After neutralization, the cells were suspended in Tyrode's buffer (138 mM NaCl, 2.9 mM KCl, 12 mM NaHCO 3 , 0.36 mM Na 2 HPO 4 , and 5.5 mM glucose, pH 7.4) containing 1.8 mM CaCl 2 and 0.49 mM MgCl 2 for adhesion studies with fibrinogen. For studies with fibronectin, cells were suspended in Hanks' balanced salt solution (136 mM NaCl, 5.3 mM KCl, 0.33 mM Na 2 HPO 4 , 0.44 mM KH 2 PO 4 , and 5.5 mM glucose, pH 7.4) with 1.8 mM CaCl 2 and 0.49 mM MgCl 2 . 24-well tissue culture plates were coated with 12.5 g/ml fibrinogen, 12.5 g/ml fibronectin, 10 g/ml anti-␣ IIb ␤ 3 antibody P2, 10 g/ml anti-␣ v ␤ 3 antibody LM609, or 2.5 mg/ml heat-treated BSA. 200 l of 5 ϫ 10 5 cells/ml were added to each well and incubated for various time points (2.5, 5, and 10 min) at 37°C in 5% CO 2 . In some experiments, cells were incubated for 30 min with 10 M cytochalasin D, 100 nM wortmannin, 20 M PD98059, 10 M U0216, or Me 2 SO (control) prior to the adhesion experiments. The unbound cells were removed by washing, and the cells bound to fibrinogen or fibronectin were lysed in ice-cold lysis buffer (15 mM HEPES, pH 7.0, 145 mM NaCl, 0.1 mM MgCl 2 , 10 mM EGTA, 1% Triton X-100, 2 mM Na 3 VO 4 , 250 g/ml 4-(2-aminoethyl)benzenesulfonyl fluoride, 15 g/ml protease inhibitors (chymostatin, antipain, pepstatin, and leupeptin), and a phosphatase inhibitor mixture (Sigma)) as previously described (12). The non-adherent cells from the BSA-coated wells were collected, diluted 1:1 in Tyrode's buffer, and centrifuged at 100 ϫ g for 5 min; and the pellet was solubilized in lysis buffer. The lysates were incubated for 30 min on ice and clarified by centrifugation at 3750 ϫ g for 30 min, and the protein concentration was determined using a Bio-Rad protein assay kit. Cells incubated with 0.5 M sorbitol or 100 nm PMA served as a positive control for MAPK activation. In some experiments, MAPK activation was assessed after monoclonal antibody clustering of the ␣ IIb ␤ 3 receptor. 5 ϫ 10 5 cells suspended in 50 l of Tyrode's buffer containing 1.8 mM CaCl 2 and 0.49 mM MgCl 2 , pH 7.4, were first incubated with 10 g/ml F(abЈ) 2 fragment from antibody 10E5 or c7E3 for 30 min at 4°C, followed by a 1:200 dilution of Fabspecific goat anti-mouse IgG for 20 min at 37°C as described in other systems (17). The cells were washed once with 200 l of Tyrode's buffer and lysed.
Adhesion of Platelets to Immobilized Fibrinogen-Blood was obtained in acid/citrate/dextrose from normal, healthy, and fasting donors of known Pl A genotype. Washed platelets were prepared as described (13), suspended in Tyrode's buffer, and allowed to recover for 2 h at 37°C. Fibrinogen (12.5 g/ml) or heat-denatured BSA was immobilized on six-well plates as described above. 1 ml containing 1 ϫ 10 8 platelets was added to each well and incubated for 15 min at 37°C in 5% CO 2 . The fibrinogen-bound platelets and the non-adherent platelets from the BSA-coated well were lysed in ice-cold lysis buffer, and the protein content was determined.
Adhesion Using a Parallel Plate Flow Chamber-Glass coverslips were coated with 200 l of 12.5 g/ml fibrinogen and incubated for 3 h in a humidified chamber. Before using each coverslip, excess fibrinogen was rinsed with Tyrode's buffer, and then the parallel plate flow chamber was assembled with the coverslip forming the base of the chamber. Measurements of cell adhesion under flow conditions were as we have previously described (18). During the experiments, the parallel plate flow chamber was mounted on an inverted-stage microscope (Eclipse TE300, Nikon Instruments, Melville, NY) equipped with a ϫ20 phase objective and a high-speed camera (Quantix Photometrics, Photometrics Ltd., Tucson, AZ) connected to a computer terminal. In most experiments, 5 ϫ 10 5 cells were perfused through the chamber for 5 min at a constant flow rate of 25 s Ϫ1 , which produces a constant wall shear stress. In some cases, cells were pretreated for 30 min with 20 M MAPK inhibitor PD98059 or 10 g/ml ML-7 and control Me 2 SO before perfusion. Four fixed fields of observation were identified, and the number of cells adhering to these regions in each experiment was counted using the Metamorph imaging system.
Clot Retraction Assay-Clot retraction was measured as we have previously described (12). Briefly, CHO cells (4 ϫ 10 6 ) were pretreated with either Me 2 SO or 20 M PD98059 for 1 h. Cells were washed once with Tyrode's buffer; resuspended in 300 l of ␣-minimal essential medium containing 28 mM CaCl 2 and 25 mM HEPES, pH 7.4; and mixed with 200 l of fibronectin-depleted plasma, 250 g of fibrinogen, and 5 g of aprotinin. Clot formation was initiated by the addition of 2.5 units of thrombin. After incubation at 37°C for varying time periods, the volume of liquid not incorporated into the clot was measured.
Platelet Secretion Studies-Secretion of ␣-granules was studied by assaying for P-selectin expression using a modification of our previously described whole blood flow cytometric assay (13). Briefly, 10 l of washed platelets (2 ϫ 10 6 ) of known Pl A genotype were incubated with 10 l of 0.2 mg/ml fluorescein isothiocyanate-labeled anti-P-selectin antibody in a total of 40 l of Tyrode's buffer containing 10 mM HEPES and activated using varying concentrations of thrombin for 2 min. Samples were fixed with 0.5% paraformaldehyde and analyzed for P-selectin expression by flow cytometry. In some experiments, platelets were incubated with Me 2 SO or 10 g/ml ML-7 before thrombin stimulation.

RESULTS
Immobilized Fibrinogen Induces Outside-in Signaling to ERK2, but Not to JNK and p38 -To determine whether integrin ␣ IIb ␤ 3 can induce outside-in signaling to MAPK, CHO cells overexpressing ␣ IIb ␤ 3 were allowed to adhere to immobilized fibrinogen or maintained in suspension over a BSA substrate. Substantially greater levels of phosphorylated ERK2 were detected in adherent Pro 33 cells compared with Leu 33 cells at 2.5, 5, and 10 min (Fig. 1A, lane 3 versus lane 2). Compared with Leu 33 cells, Pro 33 cells exhibited an ϳ10-fold increase in phosphorylated ERK2 at 2.5 min and an ϳ5-fold increase at 5 and 10 min (Fig. 1B). The levels of total ERK in Leu 33 and Pro 33 cells were equivalent and could not account for the signaling differences (Fig. 1A). ERK2 signaling was dependent upon ␣ IIb ␤ 3 and fibrinogen because 1) no phosphorylated ERK2 was detected in the vector control LK cells not expressing ␣ IIb ␤ 3 (Fig. 1A, lane 1); 2) cells maintained in suspension over the BSA substrate did not trigger activation of ERK2 (lanes 4 -6); and 3) adhesion to fibrinogen under these conditions is completely inhibited with integrelin and ␣ IIb ␤ 3 -specific functionblocking antibody 10E5 (12). A 2.5-fold increase in ERK2 phosphorylation was observed in Pro 33 cells compared with Leu 33 cells in response to soluble fibrinogen binding (data not shown).
Although CHO cell lines were generated by cell sorting and should not have been subject to clonal variation, we generated a second set of 293 cell lines to confirm the effects of the substitution of Leu with Pro at amino acid 33. Essentially the same results were obtained with the 293 cells as with the CHO cells ( Fig. 1, C and D). Densitometry revealed that compared with 293Leu 33 cells, 293Pro 33 cells demonstrated 7-fold greater phosphorylation of ERK2 at 2.5 min (data not shown). Relative to CHO cells, in 293 cells, the Pro 33 -dependent difference in ERK2 activation appeared to lessen over time (data not shown), perhaps reflecting cell-type specificities. Nevertheless, at early time points in both cell lines, the ␤ 3 Pro 33 substitution exhibited greater activation of ERK2. For the data shown in the rest of this report, we show studies with CHO cells, although similar results were obtained in 293 cells.
Under similar conditions, adhesion of Leu 33 and Pro 33 cells to fibronectin did not cause activation of ERK2 ( Fig. 2A), indicating that the differential signaling due to the Pro 33 polymorphism of integrin ␤ 3 is ligand-specific. The cells used in these experiments were fully capable of activating ERK2 because PMA and sorbitol caused robust activation in all three cell lines (Fig. 2B, lanes 1-6). The absence of ERK signaling on fibronectin might reflect trans-dominant inhibition of integrin signaling, perhaps due to the cross-talk between the overexpressed ␣ IIb ␤ 3 and the endogenous ␣ 5 ␤ 1 integrins.
To determine whether the enhanced phosphorylation of ERK2 was mediated primarily through ␣ IIb ␤ 3 , ERK2 activation was examined in cells adhering to wells coated with anti-␣ IIb ␤ 3 antibody P2, anti-␣ v ␤ 3 antibody LM609, or BSA. There was an ϳ10-fold greater phosphorylation of ERK2 in Pro 33 cells compared with Leu 33 cells that adhered to antibody P2 (Fig. 3, A  and B). No ERK2 activation was detected in LK cells (Fig. 3A,  lane 4) and in cells maintained in suspension over BSA substrate (lanes 7-9). Adhesion to LM609-coated wells did not trigger activation of ERK2 in either Leu 33 or Pro 33 cells (Fig.  3A), indicating that low levels of chimeric hamster-human ␣ v ␤ 3 expressed in these cell lines did not contribute to ERK2 signaling. We next examined whether other related MAPK members are activated by the integrin ␣ IIb ␤ 3 -fibrinogen interaction. In contrast to ERK2 activation, no phosphorylated JNK or phosphorylated p38 was observed in response to cell adhesion to fibrinogen at 2.5, 5, and 10 min (Fig. 4, A and B). However, JNK and p38 in these cells were fully activated in response to 0.5 M sorbitol treatment (Fig. 4, A and B, lanes 7-9). These studies indicate that 1) integrin ␣ IIb ␤ 3 is capable of inducing outside-in signaling to ERK2, but not to JNK or p38; and 2) compared with the Leu 33 isoform, the Pro 33 variant of ␣ IIb ␤ 3 confers early and efficient ERK2 signaling.
␣ IIb ␤ 3 Cross-linking Activates ERK2-Adhesive processes often involve multivalent receptor-ligand interaction and the clustering of integrins. We used antibody-mediated cross-linking to further assess the role of this mechanism in the Pro 33 signaling effect. Antibody cross-linking of ␣ IIb ␤ 3 resulted in greater activation of ERK2 in Pro 33 cells compared with Leu 33 cells (Fig. 5A, lane 3 versus lane 2), whereas incubation with either antibody 10E5 or the secondary antibody alone did not cause activation in either cell line (lanes 4 -9). Similar results were obtained in both CHO cells (Fig. 5B, lane 3 versus lane 2) and 293 cells (data not shown) using anti-␤ 3 antibody c7E3. Cross-linking with antibody c7E3 showed a small basal level of ERK2 phosphorylation even in vector-only (LK) cells (Fig. 5B), perhaps due to the ability of the Fab fragment from antibody c7E3 to bind other endogenous integrins on CHO cells. In these cross-linking studies, densitometry showed 1.5-3.1-fold greater ERK2 phosphorylation in Pro 33 cells compared with Leu 33 cells (data not shown).

ERK2 Activation Is Mediated by MEK and Requires Postligand Binding Events-Sequential activation of the Ras
GTPase and the kinases Raf and MEK is the best characterized pathway for activation of the ERKs (15). To determine whether activation of ERK2 in CHO cells adherent to fibrinogen was mediated through MEK, we preincubated cells in the presence and absence of the MEK inhibitors PD98059 and U0216 and studied cell interactions with fibrinogen. Adhesion of Me 2 SOtreated (control) cells to immobilized fibrinogen caused greater activation of ERK2 in Pro 33 cells than in Leu 33 cells (Fig. 6

, A, lane 3 versus lane 2; and B, lane 5 versus lane 3).
In contrast, treatment with PD98059 (Fig. 6A) or U0216 (Fig. 6B) completely prevented the induction of ERK2 activity in both Leu 33 and Pro 33 cells. Because PD98059 binds to inactive MEK and prevents Raf from phosphorylating MEK, these data demonstrate that activation of ERK in Leu 33 and Pro 33 cells is mediated through upstream MEK/Raf.
To examine whether cytoskeleton assembly has a role in ERK2 activation, cells were preincubated with cytochalasin D. Cytochalasin D completely prevented the induction of ERK2 activity in both Leu 33 and Pro 33 cells (Fig. 6C, lanes 2 and 3  versus lanes 5 and 6). Cytoskeletal reorganization in response to integrin activation is activated by lipid kinases like phosphoinositide 3-kinase (19), and we used the phosphoinositide 3-kinase inhibitor wortmannin to examine the possible effect of phosphoinositide 3-kinase on the ␣ IIb ␤ 3 -mediated activation of ERK2. Induction of ERK2 activity in both Leu 33

FIG. 3. Activation of ERK2 in CHO cells adhering to immobilized antibodies.
A, immunoblot showing phosphorylated ERK2 from the three CHO cell lines that adhered to antibody LM609 (lanes 1-3) or antibody P2 (lanes 4 -6) or that were maintained in suspension over a BSA matrix (lanes 7-9) for 5 min. The blot was probed with antiphospho-ERK antibody (pERK2; upper panel) or anti-ERK antibody (lower panel). B, densitometric quantification of ERK2 activation in cells that adhered to antibody P2 at 5 min in three different experiments. The increased ERK2 activation in Pro 33 (A2) over Leu 33 (A1) cells was significant (p ϭ 0.03) as determined by repeated measure analysis of variance. C, mean fluorescence intensity of antibody P2 binding to the three CHO cell lines as determined by flow cytometry. These signaling differences were not due to either differences in total ERK levels (A, lower panel) or differences in the surface expression of ␣ IIb ␤ 3 (C). was ablated by wortmannin (Fig. 6C, lanes 2 and 3 versus lanes  8 and 9). These results indicate that activation of ERK2 is dependent on post-ligand binding events such as actin polymerization and phosphoinositide 3-kinase signaling.
Outside-in Signaling in CHO Cells and Platelets to MLC-Compared with Leu 33 -expressing CHO cells, Pro 33 cells show a dramatically increased reorganization of the actin cytoskeleton when bound to immobilized fibrinogen (12). Because myosin and MLCK are critical for reorganization of the actin cytoskeleton and MLCK is a cytoplasmic substrate of ERK2 (10), we examined outside-in signaling in CHO cells and human platelets to MLCK. MLCK phosphorylates Thr 18 and Ser 19 on MLC, and diphosphorylated MLC can be detected by a specific antibody (20). Adhesion to immobilized fibrinogen caused greater phosphorylation of MLC in Pro 33 cells than in Leu 33 cells (Fig.  7A, lane 3 versus lane 2). Compared with Leu 33 cells, Pro 33 cells exhibited an ϳ1.8-fold increase in the levels of diphosphorylated MLC. Because MLCK activation is required for MLC phosphorylation at Thr 18 and Ser 19 , we interpret the data in Fig. 7 to mean that MLCK has been differentially activated in the Pro 33 and Leu 33 cell lines. The difference in MLCK activation between Leu 33 -and Pro 33 -expressing CHO cells was largely transient and appeared to lessen at later time points (data not shown).
The above studies were conducted exclusively in ␣ IIb ␤ 3 -expressing CHO or 293 cells. We next assessed MLCK activity in human platelets expressing the Pro 33 isoform of ␤ 3 . Compared with platelets lacking the Pro 33 form, adhesion of Pro 33 -positive platelets to fibrinogen resulted in an enhanced diphospho-rylation of MLC (Fig. 7B, lane 1 versus lane 3). No phosphorylation of MLC was detected in platelets maintained in suspension over BSA substrate. Compared with Pro 33 -negative platelets, adhesion of Pro 33 -positive platelets to fibrinogen revealed a 3.5-fold increase in the levels of diphosphorylated MLC (Fig. 7C).

Inhibition of ERK2 and/or MLC Activation Abolishes Enhanced Functional Effects in Pro 33 -expressing Platelets and
Cells-Fibrin clot retraction is a classic ␣ IIb ␤ 3 -mediated outside-in signaling function. Compared with Leu 33 cells, the Pro 33 variant exhibited a small but significant increase (p ϭ 0.02) in fibrin clot retraction at varying time points, and this increase was abolished by the MEK inhibitor PD98059 (p ϭ 0.63) (Fig. 8, A and B). PD98059 was also able to abolish the greater adhesion of Pro 33 cells to fibrinogen under shear stress in a parallel plate flow chamber (data not shown).
Because MLCK signaling modulates platelet secretion (11), we examined the functional consequence of enhanced MLCK signaling in Pro 33 platelets. Compared with Leu 33 platelets, Pro 33 -positive platelets exhibited an ϳ3-fold increase in ␣-granule secretion (as reflected in P-selectin expression) in response to 0.5 units of thrombin (p ϭ 0.04), and this increase was abolished by the MLCK inhibitor ML-7 (p ϭ 0.35) (Fig. 9,  A and B). We also tested the effect of MLCK inhibition on cell adhesion under fluid shear stress. Compared with Leu 33 cells, significantly more Pro 33 cells adhered to fibrinogen under shear stress (p ϭ 0.005) (Fig. 9C), and MLCK inhibition abolished the difference in adhesion between Leu 33 and Pro 33 cells (p ϭ 0.792 in the presence of ML-7). The enhanced adhesion of 293Pro 33 cells to fibrinogen could also be abolished by treatment with ML-7 (Fig. 9D). These data indicate that enhanced FIG. 4. Activation of JNK and p38 in CHO cells adhering to immobilized fibrinogen. CHO cells were studied, and lysates were processed as described in the legend to Fig. 1 and immunoblotted for active and total JNK (A) and p38 (B). A, immunoblot probed with anti-phospho-JNK antibody (pJNK2/pJNK1) recognizing activated JNK1 (p46) and JNK2 (p54). The same blot was stripped and reprobed with anti-JNK1 antibody. Cells treated with 0.5 M sorbitol (lanes 7-9) was included in every experiment to confirm the ability to detect phospho-JNK and to demonstrate that JNK can be activated in these cells. Data shown are representative of three experiments. B, immunoblot probed with anti-phospho-p38 (pP38) and anti-p38 antibodies. Lanes 7-9 show cells treated with 0.5 M sorbitol. FGN, fibrinogen; A1, Leu 33 cells; A2, Pro 33 cells.

FIG. 5. Activation of ERK2 in CHO cells using antibody-mediated cross-linking of integrin ␤ 3 .
Integrin ␣ IIb ␤ 3 in CHO cells was cross-linked using either antibody 10E5 (A) or c7E3 (7E3; B), followed by goat anti-mouse antibody (GAM). Total ERK levels were not different in each lane (A, lower panel). This blot is representative of three different experiments. Shown in C is the mean fluorescence intensity of antibody P2 binding to the three CHO cell lines as determined by flow cytometry. Surface expression of ␣ IIb ␤ 3 was not detectably different between the Leu 33 (A1) and Pro 33 (A2) cells and could not account for the signaling difference. pERK2, phospho-ERK2.
ERK2 and/or MLC signaling through the Pro 33 isoform of integrin ␤ 3 regulates cellular functions of ␣ IIb ␤ 3 .

DISCUSSION
Outside-in signaling is crucial for linking integrin ligation with numerous cellular processes, including adhesion, spreading, migration, and clot retraction. In this study, we used CHO and 293 cells and human platelets to evaluate the impact of the Leu 33 /Pro 33 polymorphism on ␣ IIb ␤ 3 outside-in signaling. The major findings of this study demonstrate that the Pro 33 form of ␣ IIb ␤ 3 can induce enhanced outside-in signaling via MAPK and MLCK and that these signaling pathways are indispensable for the hyperfunctional responses shown by Pro 33 platelets and cells. This study identifies a novel means for regulating integrin function through post-receptor occupancy, i.e. through ERK2 and MLCK signaling, and provides further insights regarding the molecular mechanisms responsible for the prothrombotic phenotype of a common inherited variation in integrin ␤ 3 .
Integrin-mediated ERK2 Activation-Several different approaches and two different sets of ␣ IIb ␤ 3 -expressing cell lines were used to address the impact of the Leu-to-Pro substitution at position 33 of integrin ␤ 3 on post-receptor occupancy signaling. Compared with Leu 33 cells, Pro 33 cells demonstrated substantially greater activation of ERK2 when cells bound to immobilized fibrinogen (Figs. 1 and 6). This enhanced ERK2 phosphorylation was substrate-specific (Fig. 2) and dependent upon ␣ IIb ␤ 3 (Fig. 3) and an intact actin cytoskeleton and signaling through MAPK kinase and phosphoinositide 3-kinase (Fig. 6). What are the consequences of this enhanced ERK2 phosphorylation? Integrin-mediated ERK2 activation has been most intensively studied as a regulator of gene expression and cell proliferation, but this pathway also regulates haptotactic and chemotactic cell migration (21). Cell adhesion and spreading are inhibited by dominant-negative ERK (22) and promoted by ERK activation (23). Inhibition of ERK activation blocks the formation of peripheral actin microspikes (24), and active ERK is targeted to newly forming focal adhesions after integrin engagement (25). We observed the Pro 33 isoform to enhance actin polymerization, spreading, adhesion, and clot retraction upon integrin engagement (12) and have observed a Pro 33mediated increase in haptotactic migration (26). MAPK kinase inhibition had little effect on Leu 33 cell adhesion to immobilized fibrinogen (data not shown) or clot retraction (Fig. 8), but MAPK kinase inhibition abolished the Pro 33 enhancement of these cellular functions. Thus, the enhanced ERK2 activation seen in cells expressing the Pro 33 variant of integrin ␣ IIb ␤ 3 would be predicted to have greater in vivo adhesive and migratory properties.
Integrin-mediated MLCK Activation-Human platelets and CHO cells expressing the Pro 33 variant exhibited enhanced activation of MLCK compared with Leu 33 -expressing cells upon adhering to immobilized fibrinogen (Fig. 7). Phosphorylation of MLCK is a critical regulatory step in myosin function, promoting myosin ATPase activity and increasing an actinomyosin contractile response that is involved in platelet shape change and secretion, regulation of cell migration, and polymerization of actin cables (10). This is consistent with our previous studies in which we identified 1) greater actin polymerization, adhesion, and migration of Pro 33 cells on fibrinogen compared with Leu 33 cells (12, 26) and 2) a lower threshold for platelet activation, ␣-granule release, and fibrinogen binding in Pro 33 homozygous platelets compared with Leu 33 -expressing platelets (13). Because MLCK is a substrate for ERK2, the results with PD98059 and ML-7 (Figs. 8 and 9) strongly suggest that outside-in signaling via ERK and MLCK controls the enhanced functions of adhesion, clot retraction, and secretion in Pro 33 cells and platelets. It is therefore conceivable that the enhanced ERK2/MLCK signaling in Pro 33 cells following the ligation of ␣ IIb ␤ 3 leads to a greater cytoskeletal change and favors stronger and sustained adhesion compared with Leu 33 cells. Our findings indicate that platelet physiology and signaling will be altered between the Leu 33 and Pro 33 forms of ␣ IIb ␤ 3 and that the potential prothrombotic consequences apply to a large number of individuals.
Integrin Cross-linking Enhances ERK Activation in Pro 33 Cells-Our studies show that cross-linking integrin ␣ IIb ␤ 3 enhanced activation of ERK2 in Pro 33 cells (Fig. 5) in a manner similar to cell adhesion to immobilized fibrinogen (Fig. 1). This suggests that clustering of integrin ␣ IIb ␤ 3 following adhesion to fibrinogen may underlie the enhanced activation of ERK2 in Pro 33 cells. This idea is supported by the rapid remodeling of cytoskeletal machinery in Pro 33 cells (12) and observations that identify the cytoskeletal apparatus as a key component in the process of integrin clustering (27). The requirement of intact cytoskeletal structures for ERK2 activation (Fig. 6C) is also consistent with earlier observations that cytochalasin D blocks integrin-mediated MAPK and Raf activation (16). It is likely that disruption of actin structure precludes the formation of a highly ordered cytoskeletal system that is essential for the recruitment of signaling molecules vital for activation of ERK2. Indeed, phosphoinositide 3-kinase may be one such intermediate signaling molecule because wortmannin completely abolished ERK2 activation (Fig. 6C). These results suggest that the enhanced signaling to ERK2 in Pro 33 cells is dependent on post-fibrinogen binding events, involving clustering of ␣ IIb ␤ 3 with subsequent actin rearrangement and signaling through phosphoinositide 3-kinase.
Extracellular Structure and Intracellular Signaling-Integrin cytoplasmic domains have been shown to play an important role in integrin signaling (28), and this study illustrates that residues in the extracellular region of integrin ␤ 3 can also contribute to signal transduction. Similar regulation of intracellular signals by the extracellular domain of integrin ␤ 1 has been reported (29). How could a conformational change in the extracellular domain (30) of ␤ 3 alter intracellular signaling? We are currently pursuing two major possibilities wherein the altered extracellular conformation in the Pro 33 isoform might physically 1) alter the cytoplasmic domain for a more efficient juxtaposition of ␤ 3 tails with proximal signaling molecules like Src and Syk and/or 2) induce or inhibit associations with transmembrane signaling molecules (e.g. PECAM (platelet endothelial cell adhesion molecule), Fc␥ receptor IIA, JAM1 (junctional adhesion molecule-1), etc.) that either enhance or repress, respectively, intracellular signaling. In either case, ERK2 activation could be modulated through focal adhesion kinase-dependent or -independent pathways.
ERK and MLCK Signaling in Platelets-ERK signaling is important for megakaryocyte differentiation and proplatelet formation (31,32), for the glycoprotein Ib-IX-dependent activation of platelet integrin ␣ IIb ␤ 3 (33), and for regulating the release of stored Ca 2ϩ (34). The effect of ERK signaling on platelet aggregation and secretion appears to depend on the concentration of agonists: inhibition of ERK activation blocks aggregation to low doses of collagen, arachidonic acid, U46619, and thrombin (33,35), but high concentrations of agonists can induce aggregation despite ERK inhibition (36,37). Signaling through ERK also regulates cell spreading (31,38,39). Perhaps the major effect of ERK signaling in platelets is to regulate this post-receptor occupancy process that may be distinct from agonist-induced inside-out signaling. This would be quite consistent with previous data showing no difference in fibrinogen binding to the Pro 33 and Leu 33 isoforms of ␤ 3 (13,40), but consistent differences in outside-in processes such as bleeding times, spreading, actin reorganization, and clot retraction (12,41). Moreover, signaling through MLCK is required for platelet aggregation and secretion in response to ADP (42,43). Pro 33expressing platelets showed greater activation of MLCK (Fig.  7), and this correlates with the increased aggregation and greater secretion of P-selectin in Pro 33 -positive platelets in response to ADP (13,44). Importantly, in this study, we have demonstrated that the enhanced MLCK signaling in Pro 33 platelets is required for its increased platelet reactivity, as determined by ␣-granule secretion (Fig. 9, A and B). Finally, ERK and p38 phosphorylate and activate cytoplasmic phospholipase A2 in platelets (45)(46)(47), which would release more arachidonic acid in Pro 33 -positive individuals during platelet aggregation. This hypothesis is supported by data from our laboratory (13,48,49) and others (41) showing that Pro 33expressing platelets and cell lines exhibit a greater dependence on cyclooxygenase than do cells expressing only Leu 33 .
In conclusion, we have shown that integrin ␣ IIb ␤ 3 engagement with immobilized fibrinogen induces outside-in signaling to ERK2 and MLCK and that the Leu 33 /Pro 33 polymorphism regulates the extent of this signaling. Taken in the context of known functions of ERK and MLCK, our findings support a mechanism whereby the Pro 33 variant has little effect on direct agonist-induced integrin activation, but rather enhances signaling and cell adhesive functions after ␣ IIb ␤ 3 has been engaged or cross-linked by fibrinogen. This prothrombotic phenotype may partially explain the reported arterial thrombotic risk for Pro 33 -positive individuals in some clinical epidemiology studies.